Scholarlink

Objectives <p>Develop statistical methodology for computing estimates of crime rates and characteristics from National Incident-Based Reporting System (NIBRS) data that accounts for missing data, produces estimates that are consistent across population domains and levels of geography, and can be replicated or adapted by other researchers. Apply the methodology to estimate counts and characteristics of crimes for Arizona in 2021.</p> Methods <p>Offense counts are obtained for a probability sample of non-NIBRS-reporting agencies. NIBRS records are then imputed for each incident in the offense count for each sampled agency. Standard errors are calculated using the probability sampling design and variability from multiple imputations.</p> Results <p>Estimates of Arizona violent and property crime rates from the probability sample are consistent with the Uniform Crime Reporting Program time series and are similar to estimates from the Arizona Department of Public Safety. Crime rate estimates that ignore the missing data, and estimates from “The Transition to the National Incident-Based Reporting System (NIBRS): A Comparison of 2020 and 2021 NIBRS Estimates” (Federal Bureau of Investigation <CitationRef CitationID="CR23">2022d</CitationRef>), are about 40 percent lower. Estimates from the imputed dataset of characteristics such as weapon use and victim/offender demographics account for differences between the state population and the population served by NIBRS-reporting agencies.</p> Conclusion <p>The methodology produces accurate estimates of crime rates with valid confidence intervals. Estimates of offense, victim, and offender characteristics are based on imputation models with openly stated assumptions. Future research could include a large-scale study to identify “optimal” imputation donor sets for non-NIBRS-reporting agencies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimating Crime Counts and Characteristics from NIBRS Data

  • 期刊论文

  • 发表日期：

    2026年3月12日

查看全文
  • Sharon Lohr

摘要

Objectives

Develop statistical methodology for computing estimates of crime rates and characteristics from National Incident-Based Reporting System (NIBRS) data that accounts for missing data, produces estimates that are consistent across population domains and levels of geography, and can be replicated or adapted by other researchers. Apply the methodology to estimate counts and characteristics of crimes for Arizona in 2021.

Methods

Offense counts are obtained for a probability sample of non-NIBRS-reporting agencies. NIBRS records are then imputed for each incident in the offense count for each sampled agency. Standard errors are calculated using the probability sampling design and variability from multiple imputations.

Results

Estimates of Arizona violent and property crime rates from the probability sample are consistent with the Uniform Crime Reporting Program time series and are similar to estimates from the Arizona Department of Public Safety. Crime rate estimates that ignore the missing data, and estimates from “The Transition to the National Incident-Based Reporting System (NIBRS): A Comparison of 2020 and 2021 NIBRS Estimates” (Federal Bureau of Investigation 2022d), are about 40 percent lower. Estimates from the imputed dataset of characteristics such as weapon use and victim/offender demographics account for differences between the state population and the population served by NIBRS-reporting agencies.

Conclusion

The methodology produces accurate estimates of crime rates with valid confidence intervals. Estimates of offense, victim, and offender characteristics are based on imputation models with openly stated assumptions. Future research could include a large-scale study to identify “optimal” imputation donor sets for non-NIBRS-reporting agencies.